What is embodied cognition?

What is embodied cognition?

Embodied cognition is the claim that the mind is made with a body, not merely located in one. How you perceive, remember, plan and reason depends on having hands that grasp, eyes that face forward, a spine that bends in some directions and not others, and a history of moving through the world with all of it. The brain is part of this system. It is not the whole of it.

The older picture, often called the Cartesian theater, imagined a mind sitting inside the skull watching reports from the senses and sending orders back out to the muscles. Thinking happened on the inside; the body was a vehicle. Embodied cognition pushes against this. A brain in a jar, fed the right electrical signals, would not think the way you do, because it would have no body to think with.

The phrase covers several distinct ideas, and in 2002 the psychologist Margaret Wilson sorted them into six claims that are often run together:

1. Cognition is situated. Thinking happens in a real environment, in the middle of perceiving and acting, not in an abstract space.
2. Cognition is time-pressured. The mind has to work at the speed the world moves, which rules out slow, exhaustive deliberation for most tasks.
3. We off-load cognitive work onto the environment. We count on fingers, write lists, and arrange a kitchen so it remembers for us.
4. The environment is part of the cognitive system. On the strongest reading, the boundary between mind and world is a matter of convenience.
5. Cognition is for action. Perception exists to guide movement, so what we perceive is tilted toward what we might do.
6. Off-line cognition is body based. Even when we sit still and imagine, we borrow the machinery of sensing and moving to do it.

Wilson judged the last claim the best supported and the fourth the most contested, and that balance still holds. For pain, the fifth and sixth matter most: perception is in the service of action, and imagining a movement recruits much of the same machinery as making it. Both ideas show up again when we reach treatment.

What are some examples of embodied cognition?

The strongest examples are the ones that have survived replication and have a plausible mechanism behind them. A few of the most-cited findings from the 2000s, especially studies claiming that a subtle physical cue could shift judgment or behavior, have not held up well, so this list leans on the sturdier work.

Gesture and thought

People gesture when they talk even when nobody can see them, including on the phone and in the dark. Children and adults who are allowed to gesture tend to solve certain spatial and arithmetic problems more easily than when their hands are held still. Gesture seems to be part of the thinking, a way of holding an idea in the body while the words catch up, which fits Wilson's point about off-loading cognitive work.

Posture and mood

Sitting upright versus slumped appears to nudge how people feel and how confidently they report their own state, and clinicians have long noticed that bracing and collapsing postures travel with pain. The effects in controlled studies are smaller and less reliable than popular versions of this research suggested. It is fair to say posture is one input to mood among many, not a switch.

The rubber hand

In 1998 Matthew Botvinick and Jonathan Cohen seated volunteers with one hand hidden and a rubber hand placed in view. When the experimenter stroked both hands in synchrony, people began to feel the touch in the rubber hand, and to report that it felt like their own. Asked to point to their hidden hand, they pointed toward the fake one. Within minutes, agreement between what the eyes saw and what the skin felt had moved the boundary of the body.

Tool use and the body schema

When macaque monkeys were trained to use a rake to pull in food, neurons that normally responded to objects near the hand began to respond to objects near the tip of the rake. The brain had, in effect, extended its map of the arm to the end of the tool. Anyone who has parked a car or carried a long ladder through a doorway knows the human version: the tool stops being a thing you hold and becomes, for a while, part of where you end.

What is the body schema, and how does the brain map the body?

The body schema is the brain's working model of the body for the purpose of action: where the limbs are, how long they are, what joint angles are available, how much force a movement will need. It runs mostly outside awareness. You do not consult it before reaching for a cup; it is the thing that lets you reach without consulting anything.

The body image is different. It is the body as you consciously perceive and feel about it: how it looks, what you believe about it, whether you trust it. The two usually agree, but they can come apart. Someone with chronic pain may have a body image that says 'my back is damaged' while their body schema still knows, at a level below belief, how to bend. Or the reverse: the schema itself can be distorted, so the painful limb feels larger, smaller, or harder to locate than it is.

Part of the body schema is written into the brain's somatosensory and motor cortices, where the body surface is represented as a distorted map, with the hands and face taking up far more space than the trunk. These maps are plastic. In people who have lost an arm, the region that used to represent the hand can be invaded by neighboring representations, and early work from Herta Flor's group found that the extent of that shift tracked the intensity of phantom limb pain. Later studies have complicated this picture, some finding the missing hand's representation surprisingly preserved, so the exact relationship between map changes and pain remains debated.

Lorimer Moseley and Flor reviewed this evidence in 2012 and drew the practical conclusion: if chronic pain is associated with altered cortical representations of the body, then treatments that target those representations directly are worth trying. That is the shared logic of mirror therapy, which uses a reflection to show the brain a painful limb moving freely, and graded motor imagery, which rebuilds the map in stages, from recognizing left from right, to imagining movement, to watching a mirrored limb move.

What is embodied prediction?

Embodied prediction is what you get when embodied cognition meets predictive processing. Predictive processing describes the brain as a prediction machine: it constantly guesses what the senses are about to report and corrects the guess when it is wrong. We cover that framework on its own page, predictive coding. The embodied version adds that the brain is not predicting a world out there and a body in here as separate problems. It predicts them together, because the only way it ever meets the world is through the body. Andy Clark, whose 2013 paper did much to bring predictive processing into cognitive science, put it in terms of agents rather than brains: prediction is for a creature that has to act, and acting changes what comes in next. On this view a movement is a kind of prediction fulfilled. Your brain predicts the arm will be at the cup, and the muscles make it so.

Where does pain fit? In 2017 Abby Tabor and colleagues argued that pain is best understood as embodied: a protective feeling that works by negotiating the boundaries of what actions are possible for the body right now. Pain narrows the repertoire. It makes some movements feel unavailable, and it does this before any damage has to occur, which is exactly what a good protective system should do.

Put the two together and chronic pain starts to look like a particular kind of prediction. The brain has a model of what the body can safely do. In a healthy system that model is updated every time you bend, lift, or turn without harm. In a sensitized system the model stops updating. Bending is predicted to be dangerous; the prediction produces pain and guarding; the guarding means bending never happens cleanly enough to prove the prediction wrong. The pain is real. It is also a forecast, and forecasts can be revised.

How does chronic pain change the embodied self?

If the self is partly built from the body's map and its possibilities for action, then changing the map changes the self. This is the thread that runs through the whole sense of self cluster of this series, and embodied cognition gives it a mechanism. Consider someone three years into low back pain. The body schema for the lower back has become something to steer around. The map reads as fragile there, the way an old chart marks shoals. Twisting to reach the back seat has been dropped. Picking a child up from the floor has been dropped. Standing to cook for more than ten minutes has been dropped. None of these were decisions. Each one was a small prediction of harm that was never tested, so it quietly became a fact about who this person is.

• The repertoire shrinks. Movements that are never performed fall out of the body schema, and the shrinkage itself makes the remaining movements feel riskier.
• Attention narrows to the body. A brain that predicts threat from the back monitors the back, which lowers the threshold for noticing sensations there, as our interoception page explains.
• The body image hardens. 'I have a bad back' becomes 'I am someone with a bad back', a story carried into every room.
• Fear of movement takes over the planning. What clinicians call kinesiophobia is the embodied self organizing itself around not finding out.

The self that results is smaller, and it is coherent. That is the hard part. From the inside it does not feel like a distorted model; it feels like accurate knowledge of a damaged body, and sometimes part of it is. Tissue, nerves, inflammation and disease are real inputs that need proper medical assessment, and the embodied view does not replace that assessment. What it adds is that even once the tissue has been examined and treated, the model the brain built around the injury can outlast it.

How does embodied treatment work?

If the problem is a body model that has stopped updating, then the treatment has to change the body's evidence. You cannot think your way out of a body schema. It updates on sensorimotor data: what the body does, what it feels, and what the eyes see it doing. Therapeutic neuroscience uses several ways of supplying that data.

Graded movement

The oldest approach is to perform the dropped movements again, in doses small enough that the predicted harm does not occur, and to expand from there. Each clean repetition is a prediction error in the right direction. The graded sensorimotor retraining trial for chronic back pain is a careful recent example, and pain neuroscience education is often paired with it so the person understands why a flare during the process is not a sign of injury.

Mirror and imagery feedback

When even a small movement is too threatening, graded motor imagery starts earlier in the chain, with recognizing the limb and imagining it moving, which borrows Wilson's sixth view: imagining a movement recruits much of the machinery of making it. Mirror therapy then lets the brain see the painful limb moving freely by watching the reflection of the other one. The eyes are given evidence the limb cannot yet supply.

First-person virtual bodies

Virtual reality extends the mirror. In a headset a person sees a body from the first-person perspective, and when it moves as they move, the brain tends to adopt it the way it adopted the rubber hand. That virtual body can then be shown bending a little further, or more smoothly, than the real one currently does. Research on virtual embodiment and pain suggests that changing what a person sees their body doing can change how it feels to move, and our guide to virtual embodiment therapy covers the method in detail.

Karuna's program, Karuna Virtual Embodiment Training (KVET)™, is built on this principle. Over 12 weeks at home, on a provided headset, a person does short daily VR embodiment exercises designed to create safe prediction errors around movement and to retrain the brain's body map, with a weekly one-on-one session with a pain coach and a clinician reviewing the case before the start and overseeing it throughout. Karuna has published uncontrolled outcome data from one cohort, with the numbers and their limits set out on how it works. Many people improve substantially; others do not, and that page says so.

What does this mean for you?

If you live with chronic pain, embodied cognition offers a way to understand your situation that is neither 'the tissue is damaged' nor 'it is all in your head'. Your pain is real. It is produced by a nervous system doing its protective job with a body model that has become too cautious, and body models can be updated. The work is to give your brain new evidence about what your body can do, at a pace that does not confirm the old prediction.

• Notice the dropped movements. The things you no longer do without thinking are a map of where the model has stopped updating, and they are where change starts.
• Expect the self to resist. A body that has read as fragile for years will not feel safe after one good day. The model updates on repetition, not insight alone.
• Use your eyes. Watching your body move, in a mirror, on video, or in a virtual body, is evidence your brain takes seriously, sometimes more seriously than how a movement feels.
• Get the tissue checked first. Embodied retraining begins after medical assessment, not instead of it, and it should be built with a clinician who knows chronic pain.

For clinicians, the embodied view argues for assessing the body schema and the body image alongside pain intensity: two-point discrimination, left-right judgment, how a patient draws their painful region, and what they have stopped doing. It also argues for treatment that supplies sensorimotor evidence rather than only verbal reassurance. A patient can agree that their back is strong and still move as if it were not, because the part of them that moves was not in the conversation.

Karuna offers a free consultation for people whose pain has persisted after medical evaluation, including workers' compensation referrals, where the care team confirms coverage case by case. The rest of the therapeutic neuroscience series explains the prediction machinery this page sits on.